Introduction: High left ventricular mass (LVM) is associated with excess cardiovascular morbidity and mortality. To account for differences in body size LVM is often indexed to height (HT), a power of HT, or body surface area (BSA) prior to assessment of whether LVM is high. If LVM should (i.e. physiologically) increase with greater body mass index (BMI) then indexation to HT may inappropriately identify obese individuals as having excess LVM. Conversely, BSA-indexation may underestimate the prevalence of high LVM among the obese. We sought to determine whether indexation to BSA vs HT affects prevalence of high LVM across BMI strata in a community-dwelling cohort.
Methods: We stratified Framingham Offspring study participants who underwent bSSFP CMR at 1.5T (N=1475, 47%M; 65±9y) by sex and BMI category (normal, NL=18.5 to 24.9; overweight, OW=25.0 to 29.9; obese, OB≥30 kg/m2). LVM was measured from contiguous short-axis images, then indexed to HT(m), HT1.7, HT2.7 and BSA(m2). For each indexation method we identified participants in the sex-specific top quartile (TQ) of LVM, then tabulated number (and %) of participants with TQ-LVM for each BMI category.
Results: The Table shows proportion of participants with TQ LVM by BMI category. BSA-indexation gave nearly uniform TQ-LVM status across BMI categories, i.e. the prevalence of TQ LVM/BSA was approximately 25% among NL, OW and OB strata. However, indexation to HT (or powers of HT) showed an increasing gradient of TQ LVM with greater BMI category: fewer than 10% of NL participants had TQ LVM, while approximately 40% (men) or over 50% (women) of obese participants had TQ LVM. Unindexed LVM was similar to HT-indexed LVM.
Conclusion: The association of high LVM with BMI varies by indexation method. BSA-indexation yields similar prevalence of TQ LVM across BMI strata, while HT-indexation methods show greater prevalence of TQ LVM among the obese. Further (outcomes based) work is needed to identify optimal methods for indexation of LVM.